Tumor invasion and metastasis are strongly regulated by biophysical interactions between tumor cells and the extracellular matrix (ECM). While the influence of ECM stiffness on cell migration, adhesion, and contractility has been extensively studied in two-dimensional (2D) culture, extension of these concepts to three- dimensional (3D) microenvironments characteristic of most tissues has proven extremely challenging given that manipulations normally used to vary ECM stiffness (e.g., variation of matrix and crosslink density) often concurrently alter matrix pore size (confinement), which can create steric barriers that regulate invasion speed independently of mechanics. To address this challenge, we have developed a novel matrix platform based on microfabrication of channels of defined wall stiffness and geometry that allows orthogonal variation of ECM stiffness and channel width. We have used this platform to characterize the regulation of glioblastoma cell invasion by ECM stiffness and confinement, which has led us to discover that stiff, narrow pores maximize cell invasion as a consequence of enhanced polarization of traction forces. As evidenced by this and other novel findings, this platform offers the best of both worlds with respect to experimental 2D and 3D cell migration paradigms, in that it retains the throughput, standardization, and screening power of the former while capturing key biophysical regulatory elements of the latter. With the support of this IMAT R21 award, we now propose to develop this platform as a microfluidic technology for high-throughput molecular screening and analysis. We will organize our research around three specific aims: (1) To develop an enclosed microfluidic device for the directed migration of tumor cells through channels of defined geometry and stiffness; (2) To use the platform to screen small molecule libraries for agents that slow migration in a stiffness- and confinement-dependent fashion; and (3) To relate invasion speed to gene expression in primary glioblastoma tumor initiating cells through comparative proteomic analysis. The proposed studies will address an unmet need for platforms capable of rapidly identifying drugs and genes that underlie physical microenvironmental control of tumor invasion. Ours is one of the first systematic efforts to study the roles of ECM stiffness and pore size (confinement) in regulating tumor cell invasion in 3D and to apply high-throughput molecular screening approaches to a problem in cell-ECM mechanobiology. By integrating mechanobiology, tumor stem cell biology, microfluidics, and proteomics, our work will create a valuable new discovery tool that is likely to open significant new translational opportunities for clinical oncology.